Integrating ASOS Ice Observations into the Freezing Rain Accumulation National Analysis (FRANA)

Abstract The Freezing Rain Accumulation National Analysis (FRANA) is a gridded ice accumulation analysis for nowcasting and postevent verification of freezing rain (FZRA). However, FRANA chronically underestimates the coverage of ice in weakly forced FZRA environments and can also struggle with magnitude errors especially for high-magnitude ice storms. This study explores integrating ice accumulation observations from ASOS into FRANA to ameliorate errors. An inverse distance weighting scheme is developed to bias correct FRANA and add to the footprint where ice was missed. An additional texturing algorithm is also developed to add realistic heterogeneity to new ice footprints. Case studies show that integrating observations can provide improvements for decision support, especially in areas where no FZRA was detected in FRANA. However, data-denial experiments show no statistical improvement to FRANA. This is attributed to the low spatial density of the ASOS network, which is insufficient to capture the spatial heterogeneity of ice accumulation. Local micrometeorology, terrain, and other nonmeteorological factors can cause sharp gradients in ice accumulation which further compounds the issues in conducting data-denial experiments with sparse ice observations. Sparse observations also introduce nonphysical artifacts into FRANA that degrade the meteorologically realistic footprint that FRANA originally produces. Despite the lack of statistical evidence and the introduction of nonphysical artifacts, ASOS ice observations provide information for decision support in areas where FRANA completely misses ice. These results provide the scientific background for end users to make an informed decision on how to best leverage ice observations acknowledging both the immense value and the inherent limitations of sparse observations. Significance Statement Forecasters rely on national analyses of freezing rain to fill in the gaps where there are no weather stations to observe ice accumulation. However, these analyses often miss the full extent of an ice storm, especially light freezing rain, that creates hazardous conditions. We developed an algorithm to feed direct ice measurements from airport weather sensors into the freezing rain analysis to correct these errors. While this helped fill in missing parts of an ice storm, the new algorithm can cause artifacts to show up because there are so few weather stations that observe ice. This research highlights the fundamental limits of using sparse data and helps guide the development of more reliable tools for assessing ice storms.